No. 3.] VISUAL CELLS IN VERTEBRATES. 587 
soids gives a purple effect. A glance at the double cones 
shows that this purple stain, characteristic of the ellipsoid 
in the rod, is also to be found in the ellipsoid of the near 
cone; while in the far cone the stain is like that of the single 
cones. 
Il. THE FROG. 
Preparations of frog retina fixed by corrosive-acetic present 
a feature which I have not seen in Necturus when prepared 
by the same method. This is a central core (med.) in the 
outer segments of the rods (Pl. 4, Figs. 29, 30, 31) of both 
the long red rods and the short green rods. This core 
measures one-third to one-fourth the diameter of the whole 
rod. It takes none of the stains which I have tried, with 
possibly the exception of picric acid. 
The rods otherwise seem to show the same general features 
as in Necturus. The intermediate plate, ellipsoid, and parab- 
oloid show the same reactions to stains. The paraboloid 
in the frog is not as easily distinguished as in Necturus and 
has not usually been noticed by observers, but I believe it has 
been simply overlooked on account of its small size. 
The fibrils I have not studied in this material, but, in rods 
of the frog preserved in Perenyi’s fluid I have distinguished 
superficial longitudinal fibrils (Pl. 4, Fig. 27, fbri.). The 
presence of visual-cell nuclei in a double layer, as contrasted 
with the single layer of Necturus, is a difference to be noted: 
also the presence of “green rods’ and “cone color globules,” 
which [ have not observed in Necturus. 
Ill. THE GOLDFISH. 
In the goldfish (Carassius auratus) we find an outer nuclear 
layer of four or five strata of nuclei (Pl. 3, Fig. 19). There 
are present single cones, double cones, and rods; these are 
of the type of the “green rod” in frogs with long inner 
